[Piglit] [PATCH v2 05/13] piglit-dispatch: Corrections to OpenGL API spec files.

[Chad Versace]

On 03/12/2012 03:21 PM, Paul Berry wrote:
> On 12 March 2012 14:41, Paul Berry <stereotype441 at gmail.com <mailto:stereotype441 at gmail.com>> wrote:
> 
>     This patch corrects a few minor errors in the OpenGL API spec files
>     from http://www.opengl.org/registry/:
> 
>     - ALL_ATTRIB_BITS was incorrectly defined as 0xFFFFFFFF.  The correct
>      value is 0x000FFFFF, since ALL_ATTRIB_BITS is not meant to apply to
>      MULTISAMPLE_BIT (see extension specs 3DFX_multisample,
>      ARB_multisample, EXT_multisample, and SGIS_multisample).
> 
>     - IsProgramARB and IsProgramNV were incorrectly marked as aliases of
>      each other.  This is not the case, since IsProgram is used for GLSL
>      programs, and IsProgramARB/IsProgramNV are used for assembly
>      programs.
> 
> 
> Whoops, the first sentence of this paragraph should have said "IsProgramARB and IsProgramNV were incorrectly marked as aliases of IsProgram."  Sorry for the confusion.
>  
> 
> 
>     - GenVertexArraysAPPLE was incorrectly marked as an alias for
>      "GenVertexArray".  There is no such function--the correct function
>      is "GenVertexArrays".
> 
>     - TransformFeedbackVaryingsNV was incorrectly marked as an alias for
>      TransformFeedbackVaryings.  This is incorrect--the two functions
>      interpret their arguments differently.  Similarly,
>      GetTransformFeedbackVaryingNV was incorrectly marked as an alias for
>      GetTransformFeedbackVarying.
> 
>     - StencilFuncSeparateATI was incorrectly marked as an alias for
>      StencilFuncSeparate.  This is incorrect--the two functions interpret
>      their arguments differently.
> 
>     - NV functions dealing with vertex attributes
>      (GetVertexAttrib{d,f,i}vNV, VertexAttrib{1,2,3,4}{d,f,s}{,v}NV, and
>      VertexAttrib4ub{,v}NV) were incorrectly marked as an alias for the
>      corresponding core functions.  This is incorrect--NV extensions
>      handle vertex attribute numbering differently.
> 
>     The reason these errors were not caught by the maintainers of
>     opengl.org <http://opengl.org> is likely because they don't affect the contents of the
>     glext.h header file.
>     ---
>      glapi/enumext.spec |    2 +-
>      glapi/gl.spec      |   37 ++-----------------------------------
>      2 files changed, 3 insertions(+), 36 deletions(-)


I haven't checked the extension specs myself, but trust that you did the right thing.

Acked-by: Chad Versace <chad.versace at linux.intel.com>